I posted this question before and raised it again at Brainshare but I
haven't gotten a good answer, so I thought I try again.

I have a site with BM3.7 on NetWare 6 sp3. The site connects to 2
different ISPs. One connection is to a T1, the second is to a
fractional DS3 for about 6 Mbit/sec. I have static NAT configured on
the faster interface, it's the default route. There are 12 secondary
IP addresses bound to the faster side that are NATed for application
like my SMTP server, users that connect to an SSL VPN and must have a
fixed IP, and PCAnywhere connections to a couple of hosts. I have
dynamic NAT on the slower interface and I send traffic out through it
that doesn't work through the proxy.

My problems occur when every I make a change to my NAT configuration.
After adding a new NAT mapping my server will run for a while and then
slowly start to stop routing. It's a weird failure because it doesn't
fail outright, one address at a time will stop functioning, or you can
ping an address but you can't send traffic through it. When I reboot
the server it will run fine for a couple of hours, then it will slowly
start to die again.

The only cure for this problem has been to delete the tcpip.cfg and
netinfo.cfg, reinitialize the server and reconfigure from scratch. The
server is fine until someone here forgets procedure and screws it up
again. It really makes my BorderManager installation look flaky,
instead of making the IS staff at this client look flaky which is the
case. They always come back to me and say that "It shouldn't do that.
I bet it doesn't do that for everyone." Not helping my credibility
much at all.

This has been going on for a couple of years, on NetWare 5.1 and
NetWare 6, with BM 3.6 and 3.7. I have looked very closely at both the
tcpip.cfg and the netinfo.cfg and I can find no problem with it at

Has anyone any ideas or has anyone called support with a similar

Rob Aronson, President
Perfect Networks, Inc.
Hollywod, CA USA
Using GroupWise is the only way to get the message!